Human watchkeepers under microscope
There is a need to "radically rethink the role of human watchkeepers in the digital age", according to MAIB.
The Marine Accident Investigation Branch (MAIB) recently stated that "humans can be reluctant to utilise system functions that will alert them to impending problems" following the March 2025 collision between Portuguese cargo ship Solong and the US-flagged oil tanker Stena Immaculate, which was anchored 13 miles (20km) off the East Yorkshire coast.
The MAIB report found that the ships did not have "dedicated lookouts" in what were "patchy conditions".
MAIB's comments have been echoed by maritime smart safety designer, Zelim. The Edinburgh-based company said that advanced situational awareness technology should be used as a means of helping to prevent ship collisions, reduce fatalities, and improve operational safety at sea.
Zelim CEO, Sam Mayall explained that a combination of AI-powered cameras and real-time alerting can, “help crews act faster, prevent incidents from escalating, and provide trusted records post-event.”
“Beyond preventing collisions and man-overboard incidents, these systems can also provide critical forensic data,” Mayall added.
Recommendations from IMarEST
An IMarEST working group has produced several recommendations in respect to this problem. These include the promotion of synchronised progress across technology, workforce development, and regulation as a means of ensuring cohesive implementation.
One of the most notable recommendations is to try out MASS (Maritime Autonomous Surface Ships) pilot projects before proper implementation. The working group suggested including insights from these projects, especially with respect to human-machine interface design and operator fatigue.
Other proposals covered embedding sustainability assessments into each code element to ensure alignment with UN Sustainable Development Goals, and using frameworks that reflect varying national capabilities and readiness, avoiding unintended barriers to entry.
Speaking to the BBC in August, David McFarlane, Marine Safety Consultant, said that a way forward for vessels is to augment existing technology so that a mechanism such as a Bridge Navigational Watch Alarm System could be "automatically activated" as soon as a vessel leaves a port.
However, there are challenges facing the widespread use of automation, AI and similar technology in watchkeeping.
McFarlane said that because of the volume of global marine traffic passing through the UK waters, it is likely that a proposal of this kind would need to go through the IMO. With agreement needed from the member countries, any suitable measures would most likely take between four and 10 years to come into force.
The risks that come with modern technology could also have a bearing on the effectiveness of replacing human watchkeeping with AI. In an Observer Research Foundation report, ‘AI in Maritime Surveillance: Uses, Risks, and Considerations’, author Anusha Guru pinpointed a number of potential issues with this approach.
One of the most prominent risks is the threat of cyberattacks. Guru explained that tech-minded adversaries could use technology to carry out sophisticated attacks such as, “exploiting software dependencies and manipulating the training data of these AI models.”
Guru's report also highlighted the “ethical and regulatory concerns” of AI adoption in shipping. A notable question is who takes accountability for any decisions made by unsupervised autonomous vessels, such as navigation, route planning, and collision avoidance. “With limited human oversight, assigning responsibility in the event of an error becomes complex,” argues Guru.
Tell us what you think about this article by joining the discussion on IMarEST Connect.
Image: fire-damaged Solong docked in Aberdeen, Scotland. Credit: Shutterstock.
